Introduction
Mental health is at the helm of issues pressing not only America, but the world. Despite growing community efforts to counteract this crisis, a staggering 19% of Americans experienced some form of a mental health issue in 2018. The goal of this project is to spread awareness about mental health, starting with depression among college students in the United States of America.
Method
Idea
In order to maximize our quantitative knowledge on the subject, we decided to analyze a large dataset of college student "college confessions". These posts best capture student thoughts and emotions because they are anonymous and closed within a community, usually the university itself.
Tools
First, we used selenium to web scrape thousands of college confession posts hosted by Facebook groups. Concurrently, we fed these data points to an Azure text analytics server. We then normalized the pool of data to formulate a depression index for each university.
Representation
We ran our website on Flask (a python library), and plotted the data on a United States map built by d3.js, with points representing each university. The opacity represents the university's depression index while the radius represents the school's population. This visual is now on a website hosted by an Azure server.
Usage
End users can use the map by clicking on any state (it zooms in) or by clicking on any schools bubble, bringing up that school's corresponding mental health index and most recent tweets. In order to build an uplifting and supporting community, we have implemented a live feed chat that allows users to send encouraging quotes to all other users on the website. Based on the top phrases for each school's page, users can post tweets with hashtags of these phrases on their Twitter.
What's next for Uplyft
We are looking forward not only to getting better data, but to use our natural language processing algorithm to process data from communities and high schools. During the course of our hack, one problem that we encountered was not having enough data. Going forward, we would make it a priority to get access to more data and as a result, be able to summarize and help more people. Our ultimate goal is that by using our algorithm, we can not only spread awareness about mental health, but detect suicidal and self destructive behavior from social media posts that would allow for earlier intervention.
Log in or sign up for Devpost to join the conversation.